p5.34-math-longdouble @0.250.0 has problems testing on macOS High Sierra, 10.13.6

[ballapete (Peter "Pete" Dyballa)]

--->  Testing p5.34-math-longdouble
Executing:  cd "/opt/local/var/macports/build/_opt_local_var_macports_sources_rsync.macports.org_macports_release_tarballs_ports_perl_p5-math-longdouble/p5.34-math-longdouble/work/Math-LongDouble-0.25" && /usr/bin/make test 
"/opt/local/bin/perl5.34" -MExtUtils::Command::MM -e 'cp_nonempty' -- LongDouble.bs blib/arch/auto/Math/LongDouble/LongDouble.bs 644
PERL_DL_NONLAZY=1 "/opt/local/bin/perl5.34" "-MExtUtils::Command::MM" "-MTest::Harness" "-e" "undef *Test::Harness::Switches; test_harness(0, 'blib/lib', 'blib/arch')" t/*.t


  No tests with this file. Instead, mainly for the benefit
  of the author, we (hopefully) see some diagnostics from some
  test executables that were run during the Makefile.PL stage.


CC = /usr/bin/clang
OPTMIZE = -O3
erfl & isnanl check (build): try.in:14:8: warning: implicit declaration of function 'isnanl' is invalid in C99 [-Wimplicit-function-declaration]
 ret = isnanl(op);
       ^
1 warning generated.
Undefined symbols for architecture x86_64:
  "_isnanl", referenced from:
      _main in try-8d573d.o
ld: symbol(s) not found for architecture x86_64
clang: error: linker command failed with exit code 1 (use -v to see invocation)

build output (nanl check): 
execution output (nanl check): rop set as expected

build output (nan**0 check):
try2.in:50:38: warning: format specifies type 'unsigned int' but the argument has type 'unsigned long' [-Wformat]
 printf("size of long double: %u\n", ul);
                              ~~     ^~
                              %lu
1 warning generated.

nan**0 check diagnostic: 1.0L
1.0L
1.0L
size of long double: 16

build output (signbitl check): try4.in:13:8: warning: implicit declaration of function 'signbitl' is invalid in C99 [-Wimplicit-function-declaration]
 ret = signbitl(op);
       ^
1 warning generated.
Undefined symbols for architecture x86_64:
  "_signbitl", referenced from:
      _main in try4-85c55c.o
ld: symbol(s) not found for architecture x86_64
clang: error: linker command failed with exit code 1 (use -v to see invocation)

build output (sincosl check): try5.in:18:2: warning: implicit declaration of function 'sincosl' is invalid in C99 [-Wimplicit-function-declaration]
 sincosl(op, &sine, &cosine);
 ^
1 warning generated.
Undefined symbols for architecture x86_64:
  "_sincosl", referenced from:
      _main in try5-a56b59.o
ld: symbol(s) not found for architecture x86_64
clang: error: linker command failed with exit code 1 (use -v to see invocation)


t/01config_xtra.t .......... ok

NAN_POW_BUG is NOT defined
NANL_IS_UNAVAILABLE is NOT defined
ISNANL_IS_UNAVAILABLE is defined
SIGNBITL_IS_UNAVAILABLE is defined
SINCOSL_IS_UNAVAILABLE is defined
Actual nvsize == 8
Actual long double size = 10
t/aab_basic.t .............. ok

Math::LongDouble::_have_isnanl() returns 0
Math::LongDouble::_have_nanl() returns 1
Math::LongDouble::_have_signbitl() returns 0
t/abs.t .................... ok
t/arith.t .................. ok
t/base_conversion.t ........ ok
t/bool.t ................... ok
t/cmp.t .................... ok

LD_DBL_DENORM_MIN not implemented

LD_LDBL_DENORM_MIN not implemented
t/constants.t .............. ok
t/inc_dec.t ................ ok
t/int.t .................... ok

 Skipping all tests - nvtype ('double') is not 'long double'
t/ld_bytes.t ............... ok
t/log_exp.t ................ ok

Math:LongDouble::_nan_pow_bug() returns 0
t/nan_pow.t ................ ok
t/new.t .................... ok
t/nnum.t ................... ok
t/overload.t ............... ok
t/overload_cmp_reversal.t .. ok
t/overload_switch.t ........ ok

Test::Pod version: 1.52

Pod::Simple version: 3.45
t/pod.t .................... ok
t/pow.t .................... ok

FYI:
 DBL_DIG = 15
 LDBL_DIG = 18
 Default precison = 21
t/prec.t ................... ok
t/signbit.t ................ ok
t/special.t ................ ok
t/stringification.t ........ ok

using own implementation of sincosl function
t/trig.t ................... ok
All tests successful.
Files=25, Tests=341,  2 wallclock secs ( 0.11 usr  0.04 sys +  1.01 cusr  0.17 csys =  1.33 CPU)
Result: PASS

It passes tests but the failures do not look good. Another function that has changed its name?